Method, circuit, sensor and equipment for inhibiting spectrum leakage

Abstract

The embodiment of the invention relates to the technical field of digital signal processing, and discloses a method, a circuit, a sensor and equipment for inhibiting spectrum leakage. The method is used for carrying out N-point FFT processing on a time domain data sequence with the length of Ns, N-Ns is larger than or equal to 2, and the method comprises the steps that at least [(N-Ns) / 2] zero values are supplemented before and after the time domain data sequence to obtain an N-point time domain data sequence, and [] represents rounding; performing N-point FFT processing on the N-point time domain data sequence to obtain a first frequency spectrum; acquiring a roll-off coefficient of the window function based on the number of zero values complemented by the time domain data sequence; based on the roll-off coefficient of the window function, performing dynamic windowing on the first spectrum to obtain a second spectrum; and obtaining the distance, speed and / or angle of the target based on the second frequency spectrum. At least the problem of frequency spectrum leakage of a frequency domain dynamic windowing scheme in a zero-padding FFT processing process can be improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of digital signal processing technology, and in particular to a method, circuit, sensor, and device for suppressing spectrum leakage. Background Art

[0002] When Fast Fourier Transform (FFT) is performed on data, there is usually a problem of spectrum leakage. To solve this problem, time domain windowing or frequency domain windowing can be used for suppression, but the suppression effect is not ideal. Summary of the invention

[0003] The embodiments of the present application provide a method, circuit, sensor, and device for suppressing spectrum leakage, which are at least helpful in improving the spectrum leakage problem during FFT processing in a frequency domain dynamic windowing solution.

[0004] According to some embodiments of the present application, on the one hand, the embodiments of the present application provide a method for suppressing spectrum leakage during FFT processing, for s The time domain data sequence is processed by N-point FFT, where N, N s are all positive integers, and NN s ≥2; the method comprises: adding at least [(NN s ) / 2] zero values ​​to obtain an N-point time domain data sequence, wherein [] represents rounding; performing N-point FFT processing on the N-point time domain data sequence to obtain a first spectrum; obtaining a roll-off coefficient of a window function based on the number of zero values ​​supplemented in the time domain data sequence; dynamically windowing the first spectrum based on the roll-off coefficient of the window function to obtain a second spectrum; obtaining a distance, speed and / or angle of the target based on the second spectrum.

[0005] According to some embodiments of the present application, on the other hand, the embodiments of the present application also provide a method for suppressing spectrum leakage during FFT processing, including: when a time domain signal sequence does not meet the FFT point number, adding zero values ​​at both ends of the time domain signal sequence to obtain a zero-padded time domain data sequence that meets the FFT point number; performing FFT processing on the zero-padded time domain data sequence to obtain spectrum data; and dynamically windowing the spectrum data.

[0006] According to some embodiments of the present application, on the other hand, the embodiments of the present application further provide an integrated circuit, comprising a radio frequency module, an analog signal processing module and a digital signal processing module connected in sequence; wherein the radio frequency module is used to generate a radio frequency transmit signal and receive a radio frequency receive signal; the analog signal processing module is used to down-convert the radio frequency receive signal to obtain an intermediate frequency signal; the digital signal processing module is used to perform analog-to-digital conversion on the intermediate frequency signal to obtain a time domain signal sequence, and when the time domain signal sequence does not meet the FFT point number, zero values ​​are added at both ends of the time domain signal sequence to obtain a zero-padded time domain data sequence that meets the FFT point number, FFT processing is performed on the zero-padded time domain data sequence to obtain spectrum data, and dynamic windowing is performed on the spectrum data.

[0007] According to some embodiments of the present application, on the other hand, the embodiments of the present application further provide an electromagnetic wave sensor, comprising: a carrier; an integrated circuit as described in any embodiment of the present application, which is arranged on the carrier; an antenna, which is arranged on the carrier, or the antenna and the integrated circuit are integrated into an integrated device and arranged on the carrier; wherein the integrated circuit is connected to the antenna for transmitting radio frequency transmission signals and / or receiving radio frequency reception signals.

[0008] According to some embodiments of the present application, on the other hand, the embodiments of the present application further provide a terminal device, including: a device body; and an electromagnetic wave sensor as described in any embodiment of the present application, which is arranged on the device body; wherein the electromagnetic wave sensor is used for target detection and / or communication to provide reference information for the operation of the device body.

[0009] The technical solution provided in the embodiment of the present application has at least the following advantages:

[0010] In the pair with length N s When performing N-point FFT processing on a time domain data sequence, zero values ​​are added before and after the time domain data sequence. Compared with the method of adding zeros after the time domain data sequence, this method can reduce spectrum leakage when windowing the first spectrum obtained by subsequent FFT. In this way, the result of subsequent target detection based on the second spectrum obtained by windowing will be more accurate and reliable, thereby improving the target detection effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] One or more embodiments are exemplarily described by pictures in the corresponding drawings, and these exemplified descriptions do not constitute limitations on the embodiments. Elements with the same reference numerals in the drawings represent similar elements, and unless otherwise stated, the figures in the drawings do not constitute proportional limitations.

[0012] Figure 1The process of the method for suppressing spectrum leakage in the FFT processing process provided in the embodiment of the present application is as follows: Figure 1 ;

[0013] Figure 2 is a schematic diagram of the window function;

[0014] Figure 3 It is a schematic diagram of the time domain signal corresponding to the time domain data sequence after the end is padded with zeros;

[0015] Figure 4 It is a schematic diagram of the time domain signal corresponding to the signal frequency domain windowing corresponding to the time domain data sequence after the end is padded with zeros;

[0016] Figure 5 It is a schematic diagram of the time domain signal corresponding to the time domain data sequence after front and back zero padding;

[0017] Figure 6 It is a schematic diagram of the time domain signal corresponding to the signal frequency domain windowing corresponding to the time domain data sequence after front and rear zero padding;

[0018] Figure 7 The process of the method for suppressing spectrum leakage in the FFT processing process provided in the embodiment of the present application is as follows: Figure 2 ;

[0019] Figure 8 yes Figure 7 A schematic diagram of changes in storage content of a storage space involved in the method for suppressing spectrum leakage during FFT processing provided in the illustrated embodiment;

[0020] Fig. 9 The process of the method for suppressing spectrum leakage in the FFT processing process provided in the embodiment of the present application is as follows: Figure 3 ;

[0021] Fig.10 yes Figure 8 A comparison diagram of the effective sampling time corresponding to the time domain data obtained in the method for suppressing spectrum leakage in the FFT processing process provided in the illustrated embodiment and the time domain data obtained in the prior art;

[0022] Fig.11 Schematic diagram of the window function involved in the method for suppressing spectrum leakage in the FFT processing process provided in the embodiment of the present application Figure 1 ;

[0023] Fig.12 Schematic diagram of the window function involved in the method for suppressing spectrum leakage in the FFT processing process provided in the embodiment of the present application Figure 2 ;

[0024] Fig.13 The effect comparison of the method for suppressing spectrum leakage in the FFT processing process provided in the embodiment of the present application is as follows: Figure 1 ;

[0025] Fig.14 The effect comparison of the method for suppressing spectrum leakage in the FFT processing process provided in the embodiment of the present application is as follows: Figure 2 ;

[0026] Fig.15 is a flowchart of a method for suppressing spectrum leakage during FFT processing provided in an embodiment of the present application;

[0027] Fig.16 It is a schematic diagram of the structure of the integrated circuit provided in the embodiment of the present application. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the embodiments of the present application will be described in detail below in conjunction with the accompanying drawings. However, it will be appreciated by those skilled in the art that in the embodiments of the present application, many technical details are provided in order to enable the reader to better understand the present application. However, even without these technical details and various changes and modifications based on the following embodiments, the technical scheme claimed in the present application can be implemented.

[0029] The division of the following embodiments is for the convenience of description and shall not constitute any limitation on the specific implementation of the present application. The embodiments may be combined with each other and referenced to each other without contradiction.

[0030] The following is an example of millimeter wave radar signal processing to explain the distance. That is, after receiving the echo signal, the radar will perform analog-to-digital conversion on the received echo signal to obtain time domain data, and then perform digital signal processing on the time domain data, such as range dimension Fast Fourier Transform (Fast Fourier Transform, FFT), Doppler dimension (Doppler) FFT, non-coherent accumulation (Non-coherent Combing), constant false alarm rate (Constant False-Alarm Rate, CFAR), direction of arrival (Direction of Arrival, DOA) estimation, etc., to achieve target detection. That is, there are multiple FFT processing processes.

[0031] When performing FFT processing, there is usually a problem of spectrum leakage. In order to solve this problem, one solution that can be used is to perform time domain windowing, which can suppress spectrum leakage and side lobes to a certain extent, but windowing will cause the main lobe to be broadened, which will reduce the resolution of the spectrum. Another solution is to add windows in the frequency domain, such as using a dynamic windowing scheme, such as using spatially variant apodization (SVA) to perform frequency domain windowing, which adds an optimized raised cosine window to different frequency points in the frequency domain, so that compared with the aforementioned time domain windowing scheme, it can achieve an improvement in spectrum resolution under the condition of similar side lobe suppression.

[0032] However, when performing FFT processing, in order to improve the effect of FFT, time domain data sequences that do not meet the FFT point requirement are usually padded with zeros at the end of the sequence, which in turn causes the SVA algorithm's suppression performance of frequency domain sidelobes to deteriorate. That is, zero padding causes a mutation in the time domain, which in turn leads to larger frequency domain sidelobes, resulting in poor spectrum leakage suppression during FFT processing in the frequency domain dynamic windowing scheme.

[0033] On the one hand, the embodiment of the present application provides a method for suppressing spectrum leakage in the FFT processing process, which can be used for s The time domain data sequence is processed by N-point FFT, where N, N s are all positive integers, and NN s ≥2. In some embodiments, the process of the method for suppressing spectrum leakage during FFT processing is as follows Figure 1 As shown, the following steps may be included:

[0034] Step 101: fill the front and back of the time domain data sequence with at least [(NN s ) / 2] zero values ​​to obtain an N-point time domain data sequence, where [] represents rounding.

[0035] Step 102: Perform N-point FFT processing on the N-point time domain data sequence to obtain a first spectrum.

[0036] Step 103: Obtain a roll-off coefficient of the window function based on the number of zero values ​​padded in the time domain data sequence.

[0037] Step 104: dynamically add a window to the first spectrum based on the roll-off coefficient of the window function to obtain a second spectrum.

[0038] Step 105: Acquire the distance, speed and / or angle of the target based on the second frequency spectrum.

[0039] In this embodiment, when the length is N s When performing N-point FFT processing on a time domain data sequence, at least [(NNs ) / 2] zero values, compared with the method of padding zeros after the time domain data sequence, can make the spectrum leakage less when the first spectrum obtained by FFT is subsequently windowed. In this way, the result of subsequent target detection based on the second spectrum obtained by windowing will be more accurate and reliable, thereby improving the effect of target detection.

[0040] In order to facilitate those skilled in the art to better understand the effects of the above embodiments, Figure 2-6 Herein, Figure 2-6 In the figure, the horizontal axis is the sampling number, and the vertical axis is the amplitude of the signal corresponding to the time domain data.

[0041] When the adopted windowing scheme is the frequency domain dynamic windowing scheme, assuming that the shape of the window is Figure 2 As shown. For the method of padding the end of the time domain data sequence with zeros, the signal corresponding to the time domain data sequence after padding with zeros will be as follows Figure 3 As shown in the figure, the time domain data of sequence number X1 and its subsequent sequences are all 0. Figure 3 The time domain data shown is added in the frequency domain Figure 2 After the window shown, the result corresponds to Figure 4 The time domain data shown, specifically, Figure 4 In the example, the time domain data at sequence number X1 decreases by f(X1-1) relative to the previous moment, where f(x) is Figure 2 The amplitude of the window at the sequence number x shown in FIG. 1 is the amplitude of the window at the sequence number x shown in FIG. 1. For the method of adding the same number of zero values ​​at both ends of the time domain data sequence, the time domain data corresponding to the time domain data sequence after zero filling will be as follows Figure 5 As shown in the figure, the amplitude of the time domain data of sequence number X2 and subsequent sequences is 0. Figure 5 The time domain data shown is added in the frequency domain Figure 2 After the window shown, the result obtained corresponds to Figure 6 The time domain data shown, specifically, Figure 6 In the example, the time domain data at sequence number X2 decreases by f(X2-1) relative to the previous moment, where f(x) is Figure 2 The amplitude of the window at the position x is shown. Figure 5 The signal shown is Figure 3 The signals shown have the same non-zero amplitude signal portion, while Figure 5 The signal shown has an additional portion corresponding to the leading zero before the serial number X2, so X2-1>X1-1. Figure 2The shape of the window shown, serial number (X1-1)-serial number (X2-1), the signal amplitude is constantly decreasing, therefore, f(X1-1)>f(X2-1). That is, after adding a window in the frequency domain, when the same number of zero values ​​are added at both ends of the time domain data sequence, the end of the spectrum changes from f(X2-1) to 0, and when the end of the time domain data sequence is filled with zeros, the end of the spectrum changes from f(X1-1) to 0. When the same number of zero values ​​are added at both ends of the time domain data sequence, the degree of steep drop of the spectrum is reduced, therefore, the spectrum leakage will be smaller.

[0042] In order to help those skilled in the art better understand the method for suppressing spectrum leakage in the FFT processing process provided by the above embodiment, the steps included therein are explained below.

[0043] In step 101, the time domain data sequence is supplemented with [(NN s ) / 2] zero values, and obtain an N-point time domain data sequence. The time domain data sequence refers to the time domain data corresponding to the signal containing effective information that can be used for target detection.

[0044] by Fig.10 Take the chirp echo signal received by the Frequency Modulated Continuous Wave (FMCW) as an example. Fig.10 The unit of the horizontal axis is time, and the unit of the vertical axis is frequency. Fig.10 As shown in the figure, the signal between time t1 and time t2 in a chirp signal is valid. After sampling (or analog-to-digital conversion) of this part of the signal, the number of time domain data points obtained is N. s , that is, the length N s The time domain data corresponding to the signals outside the time t1 to time t2 do not belong to the time domain data sequence.

[0045] Specifically, in radar, Fig.10 The chirp echo signal shown will be sent to the analog-to-digital converter for sampling, and there are invalid time domain data in the sampled time domain data. Therefore, it is necessary to remove the invalid time domain data and retain the valid time domain data to obtain the time domain data sequence, or, by controlling the start and shutdown time of the analog-to-digital converter, the analog-to-digital converter works only between time t1 and time t2. Among them, the time domain data sequence is stored in a specified storage space before being sent to the corresponding module for FFT processing, and the module performing FFT processing reads data from the storage space and performs FFT processing.

[0046] Considering the length N sIn the scenario where the time domain data sequence is padded with zeros to obtain an N-point time domain data sequence (i.e., a time domain data sequence with a length of N), the total number of zero values ​​that need to be padded is NN s . In addition, the time domain data sequence is supplemented with [(NN s ) / 2] zero value constraints, so its essence is: when NN s When it is an even number, add [(NN s ) / 2] zero value; and when NN s When it is an odd number, add (NN s -1) / 2 zero values ​​and (NN s +1) / 2 zero values, for example, add (NN s -1) / 2 zero values ​​are added after the time domain data (NN s +1) / 2 zero values, or add (NN) in front of the time domain data s +1) / 2 zero values ​​are added after the time domain data (NN s -1) / 2 zero values.

[0047] As for the specific implementation of zero padding, in some embodiments, it can be implemented by means of storage. Figure 7 As shown, in the method for suppressing spectrum leakage in the FFT processing process, the time domain data sequence is supplemented with at least [(NN s ) / 2] zero values, which can be achieved by the following steps:

[0048] Step 701, obtaining a time domain data sequence.

[0049] Step 702, reset the storage space with a capacity of N to zero.

[0050] Step 703, in a storage space with a capacity of N, the time domain data sequence is sequentially written from the t1+1th position.

[0051] Among them, NN s When is an even number, NN s When is an odd number, When NN s is an odd number and When NN s is an odd number and When , the number of zero values ​​padded in the front of the time domain data sequence is 1 less than the number of zero values ​​padded in the back.

[0052] Accordingly, the N-point FFT processing of the zero-filled time domain data sequence can be achieved through the following steps:

[0053] Step 704, read all data stored in the storage space with a capacity of N and perform N-point FFT processing.

[0054] The acquisition of the time domain data sequence provided in step 701 is roughly the same as the acquisition of the time domain data sequence in the aforementioned existing target detection process, and will not be repeated here. In addition, before and after the execution of step 703, the data stored in the storage space will change as follows: Figure 8 (In NN s As an example, from the first (NN s ) / 2+1 position to (N+N s ) / 2 positions are changed from 0 to the time domain data in the time domain data sequence, and the rest remain unchanged, where fi represents the i-th time domain data.

[0055] In this way, by setting the storage space to zero before writing, the data in it are all 0, so that after the time domain data sequence is subsequently written in sequence from the position t1+1, the storage space from the 1st position to the t2th position remains 0, and from the t2+1th position to the t2+Nth position remains 0. s The position is the time domain data in the time domain data sequence, starting from the t2+Nth s +1 position to the Nth position are still kept as 0, so that the data stored in the storage space is the time domain data contained in the N-point time domain data sequence after zero filling. That is, it is realized by means of storage, only the data storage method is changed, and no digital signal processing method is required. It can be well integrated with data processing in other scenarios, so that the radar can adapt to different scene requirements through the same digital signal processing method, without switching the digital signal processing method, and has low difficulty in implementation and wide application scenarios.

[0056] For cases where data processing is used, such as Fig. 9 As shown, in the method for suppressing spectrum leakage in the FFT processing process, the time domain data sequence is supplemented with at least [(NN s ) / 2] zero values, which can be achieved by the following steps:

[0057] Step 901, controlling the start and shut down time of the analog-to-digital converter, so as to complete the acquisition of t2 time domain data before starting to acquire the time domain data sequence, and to complete the acquisition of (NN) after finishing the acquisition of the time domain data sequence. s -t2) time domain data collection.

[0058] Among them, NN s When is an even number, NN s When is an odd number, When NNs is an odd number and When NN s is an odd number and When , the number of zero values ​​padded in the front of the time domain data sequence is 1 less than the number of zero values ​​padded in the back.

[0059] Step 902: t2 time domain data before the time domain data sequence and (NN) after the time domain data sequence in the data output by the analog-to-digital converter are converted into s -t2) time domain data are set to zero.

[0060] The acquisition of the time domain data in step 901 is different from the aforementioned step 701 and the aforementioned method of acquiring the time domain data by the existing radar. In the aforementioned step 701 and the aforementioned method of acquiring the time domain data by the existing radar, only the time domain data contained in the time domain data sequence needs to be acquired, while in step 901, the t2 time domain data before the time domain data sequence and (NN) after the time domain data sequence are additionally reserved by controlling the start time and the shutdown time of the analog-to-digital converter. s -t2) time domain data. Specifically, Fig.10 Taking the chirp echo signal shown in the figure as an example, in the aforementioned step 701 and the aforementioned method of obtaining time domain data by the existing radar, only the time domain data corresponding to the signal between time t1 and time t2 are retained, while in step 901, the time domain data corresponding to the time between (t1-δt1) and time (t2+δt2) are retained, δt1, δt2>0, t2 time domain data can be sampled in the time length of δt1, and (NN) can be sampled in the time length of δt2. a That is to say, the time domain data corresponding to the expected time domain number time period to be retained is only an example and does not represent the actual number of time domain data.

[0061] For step 902, it can be understood that the t2 time domain data before the time domain data sequence and the ((NN s -t2) time domain data may not be 0, therefore, the time domain data outside the time domain data sequence needs to be set to zero.

[0062] In this way, zero padding is achieved through data processing, and the implementation is simple and clear with low difficulty.

[0063] The zeroing of the time domain data involved in the above embodiment can be implemented by a filter, wherein the filter is implemented by the following expression:

[0064]

[0065] Among them, x iis the i-th input data of the filter, y i is the i-th output data of the filter.

[0066] In step 102, N-point FFT processing is performed on the zero-filled time domain data sequence to obtain a first spectrum. This embodiment does not limit the implementation of N-point FFT, such as it can be implemented by a butterfly algorithm, etc., which will not be described in detail here.

[0067] In step 103, the roll-off coefficient of the window function is obtained based on the number of zero values ​​padded in the time domain data sequence. It should be noted that the window function in this embodiment may be a raised cosine window function.

[0068] In some cases, the roll-off coefficient of the window function is obtained based on the number of zero values ​​padded in the time domain data sequence, which can be achieved by the following expression:

[0069]

[0070] min=|X w (k)| 2 ,

[0071] st0≤α≤1,

[0072]

[0073] Where α is the roll-off coefficient of the window function, N is the number of FFT points, n is the sequence number of the point in the time domain, X(k) is the kth time domain data, conj() represents the operation of finding the complex conjugate, and Real() represents the operation of taking the real part of the complex number.

[0074] At this time, the window function is the raised cosine window function given by the SVA algorithm. Therefore, the existing algorithm can be used for windowing without additional development and configuration, and the implementation difficulty is low.

[0075] Considering that the time domain data processed by N-point FFT still contains zero values ​​at both ends, and based on the roll-off coefficient constraint of 0≤α≤1 given in the existing raised cosine window function, the shape of the raised cosine window function will be as follows: Fig.11 The windows shown are all located above the horizontal axis. Therefore, after adding a window in the frequency domain, a steep drop will still occur at both ends, that is, although the side lobes will be suppressed, they are still obvious.

[0076] Based on this, in some examples, the roll-off factor constraint can be relaxed so that the shape of the raised cosine window function (here only the raised cosine window is taken as an example, and other window functions can also achieve similar effects by relaxing the constraint) will be as follows: Fig.12As shown, the windows are allowed to be located below the horizontal axis. After the frequency domain is windowed, the zero-value time domain data at both ends will be smoothly filtered and removed by the raised cosine window function, so that the side lobes will be further suppressed and the robustness to noise will be further improved.

[0077] In order to help those skilled in the art better understand the relaxation of constraints, an example is given below.

[0078] In some examples, obtaining the roll-off coefficient of the window function based on the number of zero values ​​padded in the time domain data sequence can be achieved in the following way: obtaining the upper bound of the roll-off coefficient of the window function based on the number of zero values ​​padded in the time domain data sequence; obtaining the dynamic windowing roll-off coefficient of each point in the first spectrum based on the first spectrum and the upper bound of the roll-off coefficient of the window function.

[0079] For example, the upper bound of the roll-off factor of the window function is obtained by the following expression:

[0080]

[0081] Among them, α up It is the upper bound of the roll-off factor of the window function.

[0082] By further finding an upper bound that is compatible with the amount of zero padding in the range [0, 1] specified by the roll-off coefficient, the window will not be over-adjusted due to the influence of values ​​that partially deviate from the general case, thereby achieving a better windowing effect.

[0083] For example, on the basis of constructing the upper bound, the lower bound is further constrained, for example, to avoid negative values ​​that exceed the [0, 1] range specified by the roll-off coefficient. That is, the roll-off coefficient of the dynamic windowing of each point in the first spectrum is determined by the following expression:

[0084]

[0085] Among them, α i is the roll-off coefficient of the window function corresponding to the i-th data in the first spectrum, Y i is the i-th data in the first spectrum, conj() represents the operation of obtaining the complex conjugate, and Real() represents the operation of obtaining the real part of the complex number.

[0086] In this way, by providing a relaxed constraint relationship for the expression for determining the roll-off coefficient of the window function, the value of the roll-off coefficient of the window function can be adjusted, thereby achieving the purpose of further improving the sidelobe suppression effect and improving the robustness.

[0087] To facilitate understanding of the above example to further improve the sidelobe suppression effect, the following will combine N sThe spectrum diagram obtained by using different zero-padding methods and window function determination methods when N=412 and N=512.

[0088] like Fig.13 As shown, N s =412, N=512, after normalization based on the target peak, different schemes are used for simulation. The spectrum obtained by SVA algorithm processing with zero padding at the end of the time domain data is as follows Fig.13 As shown by the dotted line in the figure, the spectrum obtained by adding zeros to the end of the time domain data and performing SVA algorithm combined with Hanning window function processing is as follows Fig.13 As shown in the line formed by the points shown in the figure, the spectrum obtained by filling zeros before and after the time domain data and performing SVA algorithm processing is as follows Fig.13 As shown in the solid line, the spectrum obtained by filling zeros before and after the time domain data and performing the SVA algorithm combined with the relaxed roll-off coefficient constraint is as follows Fig.13 As shown by the dotted line. Fig.13 It can be clearly seen that the sidelobes of the SVA algorithm processing with zero padding at the front and back of the time domain data are smaller than those of the SVA algorithm processing with zero padding only at the end of the time domain data and the SVA algorithm processing combined with the Hanning window function processing with zero padding only at the end of the time domain data. Compared with the SVA algorithm processing with zero padding at the front and back of the time domain data, the sidelobes of the SVA algorithm processing combined with the relaxed roll-off coefficient constraint processing are further reduced.

[0089] The influence of noise is further considered. At this time, after considering the thermal noise, the median of the frequency dimension is used as the noise floor estimate, and normalized according to the noise floor. The simulation results are as follows: Fig.14 As shown. Fig.13 and Fig.14 It is not difficult to see that the target loss is smaller and the robustness is better when the SVA algorithm is processed by padding the time domain data with zeros at the front and back, compared with the SVA algorithm processing with zeros only at the end of the time domain data and the SVA algorithm combined with the Hanning window function processing with zeros only at the end of the time domain data, and its advantages increase with the increase of the number of zeros. And the target loss is further reduced and the robustness is better when the SVA algorithm is processed by padding the time domain data with zeros at the front and back combined with the relaxed roll-off coefficient constraint, compared with the SVA algorithm processing with zeros at the front and back of the time domain data.

[0090] Of course, the above expressions and value ranges are merely provided for illustrative purposes. In some embodiments, other value ranges or expressions may also be used, which will not be described in detail here.

[0091] Step 104: dynamically add a window to the first spectrum based on the roll-off coefficient of the window function to obtain a second spectrum. In this embodiment, the second spectrum can be represented by the following expression:

[0092]

[0093] Among them, α is the roll-off coefficient of the window function, Y i is the i-th data in the first spectrum, Y w(i) is the i-th data in the second spectrum.

[0094] It should be noted that in the present embodiment, dynamic windowing can be achieved through a filter. In some examples, dynamic windowing of the first spectrum is performed based on the roll-off coefficient of the window function. This can be achieved in the following ways: by configuring the port, the roll-off coefficient of the window function is configured for the filter; and by using the filter, the spectrum data is windowed.

[0095] Step 105, based on the second spectrum, obtain the distance, speed and / or angle of the target. Wherein, the angle may include a pitch angle and / or an azimuth angle. In the present embodiment, obtaining the distance, speed and / or angle of the target is achieved by target detection, but the present embodiment does not limit the specific algorithm of target detection, for example, coherent accumulation, constant false alarm rate (CFAR), direction of arrival (DOA) estimation and other processing can be adopted, which will not be described one by one here.

[0096] Of course, the above is only an exemplary description provided for the scenario of target detection by radar. It is understandable that in other cases, there is still the problem of spectrum leakage in the FFT processing process in the scenario of target detection by radar as described in the above embodiment, for example, FFT processing in the synthetic aperture radar (SAR) algorithm, and FFT processing in audio and video analysis. That is, the method provided in the above embodiment can also be extended to other scenarios.

[0097] refer to Figure 2-Figure 6 , and the corresponding explanations mentioned above show that as long as the time domain data is padded with zeros before and after, the end of the spectrum will change from a smaller value (relative to the zero padded at the end) to 0. When the number of zero values ​​padded at both ends is roughly equal, the steep drop of the spectrum at both ends can be minimized as a whole, thereby achieving the best sidelobe suppression effect. However, this does not mean that when the number of zero padded at both ends is not equal, the corresponding spectrum leakage suppression effect cannot be achieved. Instead, as long as the compensation is made before and after the time domain data sequence, a certain spectrum leakage suppression effect will be brought about. As the number of zero padded at both ends approaches the same, the suppression effect will become better and better.

[0098] Based on this, the embodiment of the present application also provides a method for suppressing spectrum leakage during FFT processing, such as Fig.15 As shown, the following steps are included:

[0099] Step 1501: When the time domain signal sequence does not meet the FFT point number, zero values ​​are added to both ends of the time domain signal sequence to obtain a zero-filled time domain data sequence that meets the FFT point number.

[0100] Step 1502: Perform FFT processing on the zero-filled time domain data sequence to obtain spectrum data.

[0101] Step 1503: dynamically add windows to the spectrum data.

[0102] In this way, similar to the above-mentioned embodiment, by padding the front and back zeros, the consistency effect on the side lobes can be improved compared with the method of padding the end zeros, thereby more effectively suppressing spectrum leakage.

[0103] In step 1501, when the time domain signal sequence does not meet the number of FFT points, zero values ​​are added at both ends of the time domain signal sequence to obtain a zero-filled time domain data sequence that meets the number of FFT points. It should be noted that this embodiment does not limit the number of zero values ​​added before and after. It can be as described in the above embodiment, the same number of zero values ​​can be added before and after, or different numbers of zero values ​​can be added, for example, t3 zero values ​​are compensated in the front, and N-t3-N zero values ​​are added in the back. s zero values, t3=1,2,3,……,NN s -1. It can be understood that when the shape of the window function is Fig.12 When the symmetrical shape shown is Figure 2-6 From the analysis, it can be seen that compared with the existing method of padding zero at the end, as long as zero is padded at the front and back, no matter how many zero values ​​are compensated in the front, the maximum steep drop degree of the time domain data sequence at both ends after zero padding will be reduced, that is, the spectrum suppression effect can be improved. Fig.12 In the case of the symmetrical window function shown in the figure, the same number of zero values ​​are added to the front and back, which minimizes the maximum steepness of the time domain data sequence at both ends after zero filling, thereby suppressing the leakage of frequency to the maximum extent under the zero filling scheme. When the shape of the window function is asymmetric, the number of zero values ​​added to the front and back can be determined according to the shape of the window function to suppress the leakage of frequency to the maximum extent. Therefore, the number of zero values ​​for front and back compensation can be configured according to the needs.

[0104] In some examples, when a window function with a symmetrical window shape such as a raised cosine window is used for windowing, the two ends of the time domain signal sequence are supplemented with zero values, which can be achieved by supplementing a corresponding number of zero values ​​at both ends of the time domain signal sequence. Accordingly, dynamic windowing of the spectrum data can be achieved by dynamically windowing the spectrum data according to the raised cosine window function. In this way, based on the solution provided by the present application, the suppression of spectrum leakage can be achieved to the greatest extent.

[0105] Among them, the meaning of equivalent quantity is: if (NN s ) is an even number, the number of zero values ​​added before and after the time domain data sequence is (NN s ) / 2; if (NN s ) is an even number, the number of zero values ​​added before and after the time domain data sequence is (NN s +1) / 2, and the other is (NN s -1) / 2,N s is the length of the time series data sequence, and N is the number of FFT points.

[0106] As for the implementation method of zero padding, similar to the above-mentioned embodiment, it can be implemented by means of storage or by data processing.

[0107] In the case of storage implementation, the addition of zero values ​​at both ends of the time domain signal sequence can be implemented in the following manner: obtaining the time domain data sequence; setting the storage space to zero; and sequentially writing the time domain data sequence to the middle position in the storage space after setting the zeros. Correspondingly, performing FFT processing on the zero-filled time domain data sequence can be implemented in the following manner: reading the data in the storage space and performing FFT processing.

[0108] When implemented by data processing, the addition of zero values ​​at both ends of the time domain signal sequence can be implemented in the following manner: according to the echo signal, the time domain data sequence and the time domain data before and after the time domain data sequence are obtained; the time domain data before and after the time domain data sequence are set to zero. In some cases, according to the echo signal, the acquisition of the time domain data sequence and the time domain data before and after the time domain data sequence can be implemented in the following manner: controlling the start time and the shutdown time of the analog-to-digital converter so that the analog-to-digital converter obtains the time domain data sequence and the time domain data before and after the time domain data sequence by sampling the echo signal. Of course, in some cases, according to the echo signal, the acquisition of the time domain data sequence and the time domain data before and after the time domain data sequence can also be implemented by controlling other devices on the receiving link, etc., which will not be repeated here.

[0109] It should be noted that, in the aforementioned embodiments, the specific implementation of zero padding by means of storage and data processing has been described, and will not be repeated here.

[0110] In step 1502, the zero-filled time domain data sequence is subjected to FFT processing to obtain spectrum data, which is substantially the same as the aforementioned step 102 and will not be described in detail here.

[0111] In step 1503, the spectrum data is dynamically windowed. Similar to the above-mentioned embodiment, the roll-off coefficient of the window function is first determined, and then the spectrum data obtained by FFT processing is windowed based on the window function corresponding to the determined roll-off coefficient. That is, before the spectrum data is dynamically windowed according to the raised cosine window function, the method further includes: determining the roll-off coefficient of the raised cosine window function according to a preset constraint relationship; generating a raised cosine window function according to the roll-off coefficient of the currently determined raised cosine window function; wherein the preset constraint relationship is:

[0112]

[0113] Among them, α i is the roll-off coefficient of the raised cosine window function corresponding to the i-th data in the first spectrum, Y i is the i-th data in the spectrum data, N is the number of FFT points, N s is the length of the time series data sequence, conj() represents the operation of finding the complex conjugate, and Real() represents the operation of taking the real part of the complex number.

[0114] Of course, the above is only an expression for determining the roll-off coefficient of the improved window function. In other examples, the roll-off coefficient of the window function can also be determined by other expressions, for example, the expression for determining the roll-off coefficient of the window function provided by the SVA algorithm described in the aforementioned embodiment, and the expression for determining the roll-off coefficient of the window function by the SVA algorithm combined with a relaxed roll-off coefficient constraint, etc.

[0115] The step division of the above methods is only for the purpose of clear description. When implemented, they can be combined into one step or some steps can be split and decomposed into multiple steps. As long as they include the same logical relationship, they are all within the scope of protection of this patent; adding insignificant modifications to the algorithm or process or introducing insignificant designs without changing the core design of the algorithm and process are all within the scope of protection of this patent.

[0116] Another embodiment of the present application relates to an integrated circuit, such as Fig.16 As shown, it includes a radio frequency module 1601, an analog signal processing module 1602 and a digital signal processing module 1603 which are connected in sequence.

[0117] The RF module 1601 is used to generate RF transmission signals and receive RF reception signals.

[0118] The analog signal processing module 1602 is used to perform frequency reduction processing on the RF received signal to obtain an intermediate frequency signal.

[0119] The digital signal processing module 1603 is used to perform analog-to-digital conversion on the intermediate frequency signal to obtain a digital signal, and when the time domain signal sequence does not meet the FFT point number, add zero values ​​at both ends of the time domain signal sequence to obtain a zero-padded time domain data sequence that meets the FFT point number, perform FFT processing on the zero-padded time domain data sequence to obtain spectrum data, and dynamically window the spectrum data.

[0120] Furthermore, in some embodiments, the digital signal processing module 1603 is further configured to perform analog-to-digital conversion on the intermediate frequency signal to obtain a time domain data sequence, and to fill the front and back of the time domain data sequence with at least [(NN s ) / 2] zero values ​​to obtain an N-point time domain data sequence, wherein [] represents rounding; performing N-point FFT processing on the N-point time domain data sequence to obtain a first spectrum; obtaining a roll-off coefficient of a window function based on the number of zero values ​​supplemented in the time domain data sequence; dynamically adding a window to the first spectrum based on the roll-off coefficient of the window function to obtain a second spectrum; obtaining the distance, speed and / or angle of the target based on the second spectrum to achieve the N-point time domain data sequence. s N-point FFT processing of the time domain data sequence.

[0121] Optionally, the integrated circuit may further include a data processing module, and the data processing module is used to process digital signals to achieve target detection and / or wireless communication.

[0122] Optionally, the integrated circuit may be a millimeter wave chip.

[0123] Optionally, the radio frequency receiving signal is an echo signal formed when the radio frequency transmitting signal is emitted and / or scattered by the target, and the integrated circuit is a sensor chip.

[0124] Optionally, the integrated circuit may be a millimeter wave chip.

[0125] Optionally, the radio frequency receiving signal is an echo signal formed when the radio frequency transmitting signal is emitted and / or scattered by the target, and the integrated circuit is a sensor chip.

[0126] It is not difficult to find that the above embodiment is a circuit embodiment corresponding to the method embodiment, and the above embodiment can be implemented in conjunction with the method embodiment. The relevant technical details mentioned in the method embodiment are still valid in the above embodiment, and in order to reduce repetition, they are not repeated here. Accordingly, the relevant technical details mentioned in the above embodiment can also be applied in the method embodiment.

[0127] It is worth mentioning that each module involved in the above embodiments can be a physical unit, a part of a physical unit, or a combination of multiple physical units. In addition, in order to highlight the innovative part of the present application, the present embodiment does not introduce units that are not closely related to solving the technical problem proposed by the present application, but this does not mean that there are no other units in the present embodiment.

[0128] Another embodiment of the present application relates to an electromagnetic wave sensor, comprising: a carrier, an integrated circuit arranged on the carrier, and an antenna arranged on the carrier, or the antenna and the integrated circuit are integrated into a device arranged on the carrier. The integrated circuit is connected to the antenna and is used to process the echo signal received by the antenna. The integrated circuit is the integrated circuit provided in any of the above embodiments.

[0129] When the antenna and the integrated circuit are not integrated into one device, the integrated circuit is connected to the antenna through a first transmission line, and the first transmission line can be a printed circuit board (PCB) wiring. The carrier can be a printed circuit board PCB, such as a development board, a data acquisition board or a main board of a device, etc., which will not be described in detail here.

[0130] Since the structure of the integrated circuit included in the electromagnetic wave sensor and the working principle have been described in detail in the above embodiments, they will not be described in detail here.

[0131] The present application provides a terminal device, which may include: a device body; and an electromagnetic wave sensor as described above, which is arranged on the device body; wherein the electromagnetic wave sensor is used for target detection and / or communication to provide reference information for the operation of the device body.

[0132] In some embodiments, the electromagnetic wave sensor may be disposed outside the device body, in other embodiments, the electromagnetic wave sensor may be disposed inside the device body, and in other embodiments, the electromagnetic wave sensor may be disposed partially inside the device body and partially outside the device body. This application embodiment does not limit this, and it depends on the specific situation.

[0133] It should be noted that the electromagnetic wave sensor can realize functions such as target detection by transmitting and receiving radio signals to provide the device body with measurement information of the detected target, thereby assisting or even controlling the operation of the device body. Examples of the measurement information include: at least one of relative distance, relative speed, and relative angle.

[0134] In some embodiments, the device body may be a component or product used in fields such as transportation, consumer electronics, monitoring, in-cabin detection, and health care. For example, the device body may be intelligent transportation equipment (such as cars, motorcycles, ships, subways, trains, etc.), security equipment (such as cameras), liquid level / flow rate detection equipment, intelligent wearable devices (such as bracelets, glasses, etc.), smart home devices (such as sweeping robots, door locks, televisions, air conditioners, smart lights, etc.), various communication devices (such as mobile phones, tablet computers, etc.), and gates, intelligent traffic lights, intelligent signs, traffic cameras, and various industrialized robotic arms (or robots), etc. It may also be various instruments for detecting life characteristic parameters and various equipment equipped with the instruments, such as in-cabin detection of automobiles, indoor personnel monitoring, intelligent medical equipment, consumer electronic equipment, etc.

[0135] In some embodiments, when the above-mentioned device body is applied to an advanced driving assistance system (ADAS), the electromagnetic wave sensor as a vehicle-mounted sensor can provide the ADAS system with various functional safety guarantees such as automatic emergency braking (AEB), blind spot detection warning (Blind Spot Detection, BSD), lane change assist warning (Lane Changing Assist, LCA), and reversing assist warning (Rear Cross Traffic Alert, RCTA).

[0136] In addition, the examples mentioned in the above embodiments can be freely combined, and any combination can be understood as an embodiment. The "embodiment" or "example" appearing in various positions in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It can be understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0137] Those skilled in the art will appreciate that the above embodiments are specific embodiments for implementing the present application, and in actual applications, various changes may be made thereto in form and detail without departing from the spirit and scope of the present application.

Claims

1. A method for suppressing spectrum leakage during FFT processing, characterized in that: For the length N s The time domain data sequence is processed by N-point FFT, where N, N s are all positive integers, and NN s ≥2; the method comprising: The time domain data sequence is supplemented with at least [(NN s ) / 2] zero values, and obtain an N-point time domain data sequence, where [] indicates rounding; Performing N-point FFT processing on the N-point time domain data sequence to obtain a first spectrum; Obtaining a roll-off coefficient of a window function based on the number of zero values ​​padded in the time domain data sequence; Based on the roll-off coefficient of the window function, dynamically windowing the first spectrum to obtain a second spectrum; Based on the second frequency spectrum, the distance, speed and / or angle of the target is acquired.

2. The method for suppressing spectrum leakage during FFT processing according to claim 1, characterized in that: The time domain data sequence is supplemented with at least [(NN s ) / 2] zero values, including: Acquire the time domain data sequence; Set the storage space with capacity N to zero; In the storage space with a capacity of N, the time domain data sequence is sequentially written from the t1+1th position, where NN s When is an even number, NN s When is an odd number, or The performing N-point FFT processing on the N-point time domain data sequence comprises: All data stored in the storage space with a capacity of N are read and N-point FFT processing is performed.

3. The method for suppressing spectrum leakage in FFT processing according to claim 1, characterized in that: The time domain data sequence is supplemented with at least [(NN s ) / 2] zero values, including: Controlling the start and stop time of the analog-to-digital converter to complete the acquisition of t2 time domain data before starting to acquire the time domain data sequence, and to complete the acquisition of (NN) after finishing the acquisition of the time domain data sequence. s -t2) time domain data collection, where NN s When is an even number, NN s When is an odd number, or The t2 time domain data before the time domain data sequence and (NN) after the time domain data sequence in the data output by the analog-to-digital converter are s -t2) time domain data are set to zero.

4. The method for suppressing spectrum leakage during FFT processing according to claim 3, characterized in that: The acquired data includes t2 time domain data before the time domain data sequence and (NN s -t2) time domain data is set to zero, which is achieved by the filter shown in the following expression: Among them, x i is the i-th input data of the filter, y i is the ith output data of the filter.

5. The method for suppressing spectrum leakage in FFT processing according to any one of claims 1 to 4, characterized in that: The step of obtaining a roll-off coefficient of a window function based on the number of zero values ​​padded in the time domain data sequence comprises: Obtaining an upper bound of a roll-off coefficient of the window function based on the number of zero values ​​padded in the time domain data sequence; Based on the first spectrum and an upper bound of the roll-off coefficient of the window function, a roll-off coefficient of dynamic windowing at each point in the first spectrum is obtained.

6. The method for suppressing spectrum leakage in FFT processing according to claim 5, characterized in that: The upper bound of the roll-off coefficient of the window function is obtained based on the number of zero values ​​padded in the time domain data sequence, and is implemented by the following expression: Among them, α up is the upper bound of the roll-off coefficient of the window function; The roll-off coefficient of the dynamic windowing of each point in the first spectrum is obtained based on the upper bound of the roll-off coefficient of the first spectrum and the window function, which is implemented by the following expression: Among them, a i is the roll-off coefficient of the dynamic windowing corresponding to the i-th point in the first spectrum, Y i is the ith point in the first spectrum, conj() represents an operation of obtaining the complex conjugate, and Real() represents an operation of obtaining the real part of a complex number.

7. The method for suppressing spectrum leakage in FFT processing according to any one of claims 1 to 4, characterized in that: The dynamically windowing the first spectrum based on the roll-off coefficient of the window function includes: By configuring the port, configuring the roll-off coefficient of the window function for the filter; The filter is used to perform windowing processing on the spectrum data.

8. A method for suppressing spectrum leakage during FFT processing, characterized in that: include: When the time domain signal sequence does not meet the number of FFT points, zero values ​​are added to both ends of the time domain signal sequence to obtain a zero-filled time domain data sequence that meets the number of FFT points; Performing FFT processing on the zero-filled time domain data sequence to obtain spectrum data; Dynamically window the spectrum data.

9. The method for suppressing spectrum leakage in FFT processing according to claim 8, characterized in that: The step of adding zero values ​​to both ends of the time domain signal sequence comprises: Acquire the time domain data sequence; Zero the storage space; In the storage space after zeroing, sequentially writing the time domain data sequence into the middle position; The performing FFT processing on the zero-padded time domain data sequence comprises: Read the data in the storage space and perform FFT processing.

10. The method for suppressing spectrum leakage in FFT processing according to claim 8, characterized in that: The step of adding zero values ​​to both ends of the time domain signal sequence comprises: Acquire the time domain data sequence and time domain data before and after the time domain data sequence according to the echo signal; The time domain data before and after the time domain data sequence are set to zero.

11. The method for suppressing spectrum leakage in FFT processing according to claim 10, characterized in that: The step of acquiring the time domain data sequence and time domain data before and after the time domain data sequence according to the echo signal includes: The start-up time and the shut-down time of the analog-to-digital converter are controlled so that the analog-to-digital converter obtains the time domain data sequence and the time domain data before and after the time domain data sequence by sampling the echo signal.

12. The method for suppressing spectrum leakage in FFT processing according to any one of claims 8 to 11, characterized in that: The step of adding zero values ​​to both ends of the time domain signal sequence comprises: Supplementing equal number of zero values ​​at both ends of the time domain signal sequence; The dynamically windowing the spectrum data comprises: The spectrum data is dynamically windowed according to a raised cosine window function.

13. The method for suppressing spectrum leakage in FFT processing according to claim 12, characterized in that: If (NN s ) is an even number, the number of zero values ​​added before and after the time domain data sequence is (NN s ) / 2; If (NN s ) is an even number, the number of zero values ​​added before and after the time domain data sequence is (NN s +1) / 2, and the other is (NN s -1) / 2,N s is the length of the time series data sequence, and N is the number of FFT points.

14. The method for suppressing spectrum leakage in FFT processing according to any one of claims 8 to 11, characterized in that: Before dynamically windowing the spectrum data according to the raised cosine window function, the method further includes: Determining a roll-off coefficient of the raised cosine window function according to a preset constraint relationship; Generate the raised cosine window function according to the roll-off coefficient of the raised cosine window function currently determined; The preset constraint relationship is: Among them, α i is the roll-off coefficient of the raised cosine window function corresponding to the i-th data in the first spectrum, Y i is the i-th data in the spectrum data, N is the number of FFT points, N s is the length of the time series data sequence, conj() represents the operation of obtaining the complex conjugate, and Real() represents the operation of obtaining the real part of the complex number.

15. An integrated circuit, characterized in that: It includes a radio frequency module, an analog signal processing module and a digital signal processing module connected in sequence; Wherein, the RF module is used to generate RF transmission signals and receive RF reception signals; The analog signal processing module is used to down-convert the RF received signal to obtain an intermediate frequency signal; The digital signal processing module is used to perform analog-to-digital conversion on the intermediate frequency signal to obtain a time domain signal sequence, and when the time domain signal sequence does not meet the FFT point number, add zero values ​​at both ends of the time domain signal sequence to obtain a zero-padded time domain data sequence that meets the FFT point number, perform FFT processing on the zero-padded time domain data sequence to obtain spectrum data, and dynamically window the spectrum data.

16. The integrated circuit according to claim 15, characterized in that The digital signal processing module is also used to fill the front and back of the time domain data sequence with at least [(NN s ) / 2] zero values ​​to obtain an N-point time domain data sequence, wherein [] represents rounding; performing N-point FFT processing on the N-point time domain data sequence to obtain a first spectrum; obtaining a roll-off coefficient of a window function based on the number of zero values ​​supplemented in the time domain data sequence; dynamically adding a window to the first spectrum based on the roll-off coefficient of the window function to obtain a second spectrum; obtaining the distance, speed and / or angle of the target based on the second spectrum to achieve the N-point time domain data sequence. s N-point FFT processing of the time domain data sequence.

17. An electromagnetic wave sensor, characterized in that: include: Carrier; The integrated circuit as claimed in claim 15 or 16, arranged on a carrier; An antenna is arranged on the carrier, or the antenna and the integrated circuit are integrated into one device and arranged on the carrier; The integrated circuit is connected to the antenna and is used to transmit a radio frequency transmission signal and / or receive a radio frequency reception signal.

18. A user terminal device, characterized in that: include: Equipment body; and an electromagnetic wave sensor as claimed in claim 17 disposed on the device body; The electromagnetic wave sensor is used for target detection and / or communication to provide reference information for the operation of the device body.

Images